Percutaneous transluminal coronary angioplasty (PTCA) and coronary artery bypass grafting (CABG) are effective means for mechanical revascularization of diseased, obstructive coronary arteries. The safety and effectiveness of both PTCA and CABG have been demonstrated in subsets of patients with coronary artery disease. However, a large subset of patients with coronary artery disease who present for mechanical revascularization, patients with severe symptoms and two or three vessel disease have not been studied. In such patients, successful PTCA offers an immediate relief of symptoms with an initial hospitalization of two to three days and the ability of the patient to return to full activity upon hospital discharge. However, if PTCA is unsuccessful in such patients, particularly if it results in emergency CABG, these patients present a greater problem for operative revascularization. Furthermore, when one performs PTCA at several sites in any given patient, the chance of restenosis at one site in a patient increases proportionally as the number of sites increases. Thus, the immediate benefit of this procedure must be considered against the possible complications and future hospitalizations resulting from restenosis. CABG usually provides more complete revascularization in patients with multivessel disease, however, initial hospitalization and morbidity resulting from this operation as well as the possibility of subsequent operations detract from its initial use. Because of the complexities of the risks and benefits, both short term and long term of these two different techniques of myocardial revascularization and because there are currently no studies which would allow a comparison of the two groups, the objective of this study is a randomized comparison of PTCA and CABG in the subgroup of patients with symptomatic coronary artery disease who are not controlled on medical therapy and are candidates for both procedures. In addition to the randomized studies, all patients undergoing PTCA or CABG would be closely followed to provide complete data on these procedures when applied to a variety of patients with coronary artery disease. The primary endpoints would be death, non-fatal myocardial infarction and exercise capacity and physiology. The secondary endpoints would be duration of hospitalization, need for repeat procedures ability to return to work and full recreational activity and costs.